From 93db41483a3a46b849218945b2114326e02f69f3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Javier=20Marti=CC=81nez=20A=CC=81lvarez?= Date: Wed, 27 Mar 2024 18:26:20 +0100 Subject: [PATCH] more fixes --- private_gpt/components/ingest/readers/rdfreader.py | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/private_gpt/components/ingest/readers/rdfreader.py b/private_gpt/components/ingest/readers/rdfreader.py index 3bc3ac99..eda34d3a 100644 --- a/private_gpt/components/ingest/readers/rdfreader.py +++ b/private_gpt/components/ingest/readers/rdfreader.py @@ -10,6 +10,7 @@ Original code: https://github.com/run-llama/llama-hub """ +import logging from pathlib import Path from typing import Any @@ -18,6 +19,8 @@ from llama_index.core.schema import Document from rdflib import Graph, URIRef from rdflib.namespace import RDF, RDFS +logger = logging.getLogger(__name__) + class RDFReader(BaseReader): """RDF reader.""" @@ -46,11 +49,17 @@ class RDFReader(BaseReader): if len(labels) > 0: return labels[0].value - return None # Return None if label not found + return str(uri) def load_data(self, file: Path, extra_info: dict | None = None) -> list[Document]: """Parse file.""" - lang = extra_info["lang"] if extra_info is not None else "en" + extra_info = extra_info or {} + extra_info["graph_type"] = "rdf" + lang = ( + extra_info["lang"] + if extra_info is not None and "lang" in extra_info + else "en" + ) self.g_local = Graph() self.g_local.parse(file) @@ -62,6 +71,7 @@ class RDFReader(BaseReader): text_list = [] for s, p, o in self.g_local: + logger.debug("s=%s, p=%s, o=%s", s, p, o) if p == RDFS.label: continue